effects of components in a process model on other components are determined. The effects are then used to perform preprocessing.
|
1. A method for optimizing a process model, comprising:
determining effects of components, including actions and results, in a process model on other components to perform a structure analysis in the form of a map;
optimizing the process model based on the effects, by at least one of parallelization of components, elimination of at least one component and introduction of a checked intermediate component; and
ascertaining a range suitable for optimization based on the map.
2. A system for optimizing a process model, comprising:
a processor performing structure analysis in the form of a map by determining effects of components in a process model on other components, the components including results and actions, ascertaining a range suitable for optimization based on the map, and optimizing the process model based on the effects, by at least one of parallelization of components, elimination of at least one component and introduction of a checked intermediate component.
|
This application is based on and hereby claims priority to German Application No. 199 32 945.1 filed on Jul. 14, 1999, the contents of which are hereby incorporated by reference.
1. Field of the Invention
The invention relates to a method and system for preprocessing.
2. Description of the Related Art
Within the context of preprocessing, a process model for a technical system is expediently ascertained in one step in the system engineering. Such a process model quickly becomes confusing as the complexity of the technical system increases. Associated with this are sources of error when altering, adjusting and implementing the process model. It is also possible to ascertain a process model for an already existing technical system, with the aim of improving it. Particularly when the real technical system is used as a template for the process model, the model itself quickly becomes confusing; optimization is possible only with difficulty, with enormous complexity and with a high degree of susceptibility to error.
The object of the invention is to permit preprocessing which can be used to optimize a process model systematically and in a way which is tolerant of errors.
In this context, it may be noted that the preprocessing may advantageously be used as an input for further steps, e.g. design, whether it be redesign, adjustment, control or resetting of a technical system. One advantageous feature in this context is, among other things, processing of the data in the process model, so that, by way of example, improved operation of the technical system is then ensured.
The object is achieved by specifying a method for preprocessing in which effects of components in a process model on other components are determined. The effects are used to perform the preprocessing.
One refinement in this context is that the preprocessing is an optimization of the process model.
In particular, the process model can advantageously be ascertained using a multiplicity of components of different type each having different effects. The preprocessing is expediently performed by taking into account the effects of the components. In this context, the interplay of the components among one another is taken into account in the process model. If particular effects which need to be optimized for preprocessing are taken into account, the process model can be improved or optimized for a possibly concrete implementation in the form of a technical system.
One refinement is that optimization is effected by at least one of the following steps:
a) Parallelization of components:
b) Elimination of a component
c) Introduction of a checked intermediate component
One development is that the components are distinguished in the form of results and/or actions according to their significance. Preferably, (combinable) results and (combinable) actions alternate. Hence, the effects on a result or of a result, and vice versa, on an action or of an action, can be determined.
In this context, the following effects can be determined, in particular:
a) influence of at least one result which precedes an action;
b) influence of an action on at least one subsequent result;
c) influence of at least one action which precedes a result;
d) influence of a result on at least one subsequent action.
Effects can therefore be regarded as the influencing of a result by preceding results/actions and as the effects of a result on subsequent results/actions.
Another refinement is that the preprocessing is used to perform a structure analysis. The aim of the structure analysis is to find starting points for optimization in the process model. The process model can be conditioned in terms of its structure, in particular, with the structure analysis making it possible, on the basis of this structure, to determine starting points for optimization. Use can then expediently be made of a process model in the form of a structure representation of the actions and results (alternating with one another).
Another development is that the result of the preprocessing, in particular the structure, is used for designing a technical system. In this context, the design of the technical system may comprise redesign, adjustment, validation, optimization or control of the technical system.
The object is also achieved by providing system for preprocessing which comprises a processor unit, which processor unit is set up such that
a) effects of components in a process model on other components can be determined;
b) the effects can be used to perform the preprocessing.
The object is also achieved by providing a computer program which can be used, when loaded and executed on a processor unit, to perform the following steps:
a) effects of components in a process model on other components are determined;
b) the effects are used to perform the preprocessing.
The system is particularly suitable for carrying out the inventive method or one of its developments explained above.
These and other objects and advantages of the present invention will become more apparent and more readily appreciated from exemplary embodiments of the invention illustrated and explained below with reference to the drawings, in which:
Reference will now be made in detail to the preferred embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like reference numerals refer to like elements throughout.
Influences of Results and Actions
A result E in an (engineering) process, and hence in a process model, is influenced to a certain extent by all directly preceding results. On the other hand, a result affects all directly subsequent results.
a) box=result,
b) arrow=action
c) dashed arrow=flow of information.
A result influences a subsequent result if there is at least one connection between these results via actions or flows of information.
The influence of the i-th result on the k-th result is determined as eik, where
0≦eik≦1 i, kεN. (1)
In this context, the value 1 signifies “greatest possible influence”, and the value 0 signifies “no influence”.
In addition, each action T in an (engineering) process (process model) is influenced to a certain extent by all directly preceding results. On the other side of the coin, an action affects all directly subsequent results.
There is an influence between a result and the actions arising therefrom directly or via flows of information and also between an action and the results which result therefrom directly or via flows of information.
The influence of the i-th result on the n-th action is defined as tin, where
0≦tin≦1 iεN; nε. (2)
Similarly, the influence of the m-th action on the k-th result is defined as tmk, where
0≦tmk≦1 mε; kεN. (3)
In this context, the value 1 signifies “greatest possible influence”, and the value 0 signifies “no influence”.
By way of example, in
t2a: denotes the influence of the result E2 on the action Ta;
e23: denotes the influence of the result E2 on the result E3;
ta3: denotes the influence of the action Ta on the result E3.
Influence on a Result
BE denotes the influence on a result by preceding results and/or actions. The influence on the k-th result is defined as the sum of the influences elk of all directly preceding results plus the sum of the influences of all directly preceding actions.
BEK=ΣiEik+Σmtmk i, kεN; mε (4)
Effect of a Result
WE denotes the effect of a result on subsequent results and actions. The effect of the i-th result is defined as the sum of the influences eik on all directly subsequent results plus the sum of the influences on all directly subsequent actions.
WEi=Σkeik+Σntin i, kεN; nε. (5)
Influence on an Action
Similarly, BT denotes the influence on an action by preceding results. The influence on the n-th action is defined as the sum of the influences tin of all directly preceding results.
BTn=Σitin iεN; nε. (6)
Effect of an Action
Similarly, WT denotes the effect of an action on subsequent results. The effect of the m-th action is defined as the sum of the influences tmk on all directly subsequent results
WTm=Σktmk kεN; mε (7)
Structure Analysis
Influences of Results and Actions in the Structure Analysis
A purely structural analysis of the design of process models does not take account, in particular, of the extent of the influences of results and actions; it is merely significant whether or not an influence exists. In the equations (1) to (7), the values eik, tin and tmk therefore assume the value 1 if an influence exists.
Alternatively, various influences can be expected; in that case, the values are preferably in a range between 0 and 1 (see above definitions).
Accordingly,
influence: BE3=2+1=3
(two preceding results and one preceding action)
effect: WE3=2+2=4
(two subsequent results and two subsequent actions)
and the following is obtained for the action Ta:
influence: BTa=2
(two preceding results)
effect: WTa=1
(one subsequent effect)
Other examples are given in
Coordinate System for Classifying the Model Elements
The text below considers the results of a process model. The considerations may be given in a similar manner for actions.
For each result of a process model, the influence BE and the effect WE are ascertained on the basis of equations (4) and (5). In addition, for each of these variables, the arithmetic mean over all results and the standard deviation σ from the mean are formed.
The arithmetic mean MBE of the influences is obtained on the basis of the following:
where a denotes the number of results. The standard deviation σBE from the mean of the influences is obtained on the basis of the following:
The arithmetic mean of the effects is obtained as:
The standard deviation σBE from the mean of the influences is obtained on the basis of the following:
The characteristic quantities “influence” and “effect” and also the respective means and standard deviations can now be used to classify the results. To this end, a coordinate system is determined, on whose abscissa the influence is plotted and on whose ordinate the effect is plotted (cf.
This coordinate system shows the straight lines
BE=MBE (10)
and
WE=MWE (11)
This first produces four ranges in the 1st quadrant of the coordinate system. A fifth range is defined around the point of intersection
X(MBE; MWE) (12)
of the two straight lines based on equation (10) and equation (11). To this end, the following points are first ascertained in the coordinate system:
A(MBE−fB σBE; MWE), (13)
B(MBE; MWE+fWσWE), (14)
C(MBE+fBσBE; MWE), (15)
D(MBE; MWE−fWσWE). (16)
In this case, the factor fB stipulates the distance of points A and C from the point of intersection X, and fW stipulates the distance of points B and D from the point of intersection X.
0≦fB≦3 (17)
0≦fW≦3 (18)
give distances in the range between 0 and 3σ.
Connecting points A to B, B to C, C to D and finally D to A gives a rhombus 201 whose area defines the fifth range.
For this fifth range 201, other geometric shapes (rectangle, ellipse etc.) are also conceivable.
Significance of the ranges in the coordinate system
Each model element is sorted into the coordinate system on the basis of its values for influence and effect, and in this context is put into one of the aforementioned five ranges or on the abscissa or the ordinate of the coordinate system.
The four ranges 402, 403, 404 and 405 outside the central range 401 contain the “conspicuous results”.
The structure analysis can be used to derive the pointers illustrated below for optimizing the process model.
Neutral Results
These results are inconspicuous in terms of the structure analysis and need not be considered in any further detail in this context.
Supplying Results
These results have an intense effect on a relatively large number of directly subsequent results and actions. Errors or shortcomings in such results can therefore spread many times.
In terms of their effect, such results should be kept clear and should be reviewed as appropriate.
For optimization, the following options, in particular, are therefore considered:
These results are influenced by a relatively large number of directly preceding results and actions. The diverse effects, e.g. owing to data from different results and owing to actions from a large number of different workers, mean that such results can lead to difficulties. In particular, prompt production of a product or a technical system may be compromised, clarity of the relationships may be lost on account of the very pronounced collecting character or inconsistencies in the data may arise.
Such results are carefully reviewed, particularly with regard to a production deadline, are kept clear in terms of content and are checked for consistency after production.
For optimization, the following options are therefore considered;
These results are influenced little and have little effect. The creation of such results can have a delaying effect or may be quite superfluous. Typical results are those which lie in a sequential path whose actions can be parallelized, or, by way of example, merely the format of the data contained is converted (e.g. on account of media breakage).
For optimization, the following options are considered:
Is parallelization possible?
Is it possible to dispense with the result?
These results are influenced by a relatively large number of directly preceding results and actions. On the other side of the coin, they have an intense effect on a relatively large number of directly subsequent results and actions.
The diverse effects, e.g. owing to data from different results and owing to actions from a large number of different workers, mean that such results can be problematical. In particular, the prompt production of a system may be compromised, the clarity may be lost on account of the highly collecting character and inconsistencies in the data contained may arise. These problems are particularly critical since the broad effect means that errors or shortcomings in such results can spread many times.
The results are carefully reviewed with regard to the production deadline, for example, are kept clear in terms of influence, content and effect and are checked thoroughly after production.
For optimization, the following options are therefore considered:
Within the context of analyzing a process model and hence a technical system associated with the process model, a map as shown in
As presets, the rhombus shape is provided for the neutral range, and the value “1” is provided for the factors of the equations (17) and (18). Alterations to these presets are preferably shown in the coordinate system of
After an “Analyze” button has been selected, an analysis algorithm is started. For each result, this algorithm automatically ascertains the values for influence and effect from the graphical structure, i.e. from the links between the individual model elements in the process model. The ascertained values are allocated to the individual results in the process model. This is expediently done using system attributes provided for this purpose. In addition, the means for the influence and effect of all results and the respective standard deviations are calculated.
Presentation of the Analysis Results
The result of the analysis is shown, in particular, in the coordinate system:
After the analysis result has been shown, the factors (see equations (17) and (18)) and the shape of the neutral range can be altered. The alterations are shown in the coordinate system. In particular, the items 1, 3 and 5 in the result presentation are continuously updated in this context.
Determination/Identification of Conspicuous Model Elements
After the analysis has been performed, a button denoted “Color model elements” can be selected. This preferably opens a dialog window which provides options for ways in which each of the five analysis ranges (“neutral”, “supplying”, “collecting”, “buffering”, “critical”) and also the results which are on the ordinate (“purely supplying”) and on the abscissa (“purely collecting”) can be highlighted. Preferably, each range has a color pallet for this purpose, with a different color already having been preset for each range. Thus, for example, black is preset for results in the “neutral” range, red is preset for “critical” results etc. The results can be marked with a color in the process model by clicking on the “Color” button.
The invention has been described in detail with particular reference to preferred embodiments thereof and examples, but it will be understood that variations and modifications can be effected within the spirit and scope of the invention.
Patent | Priority | Assignee | Title |
7444312, | Aug 14 2002 | Siemens Aktiengesellschaft | Method and arrangement for designing a technical system |
Patent | Priority | Assignee | Title |
6128403, | Aug 26 1997 | Mitsubishi Denki Kabushiki Kaisha | Wafer map analysis aid system, wafer map analyzing method and wafer processing method |
6292707, | Nov 12 1998 | Northrop Grumman Systems Corporation | Integrated design and manufacturing system |
6314413, | Aug 13 1997 | OTTE, RALF | Method for controlling process events using neural network |
6741898, | Sep 29 2000 | Ford Motor Company | Method of using neutral event file for manufacturing line analysis |
6748283, | Sep 29 2000 | Ford Motor Company | Method of using neutral event specification file for manufacturing line analysis |
6871171, | Oct 19 2000 | OPTIMATA LTD | System and methods for optimized drug delivery and progression of diseased and normal cells |
DE19508474, | |||
DE19634923, | |||
DE19716928, | |||
DE4319926, | |||
EP411873, | |||
EP659996, | |||
EP877329, | |||
FR2724744, | |||
WO9712300, | |||
WO9712301, | |||
WO9824010, |
Executed on | Assignor | Assignee | Conveyance | Frame | Reel | Doc |
Jul 13 2000 | Siemens Aktiengesellschaft | (assignment on the face of the patent) | / | |||
Dec 21 2001 | KODES, RUDOLF | Siemens Aktiengesellschaft | ASSIGNMENT OF ASSIGNORS INTEREST SEE DOCUMENT FOR DETAILS | 012725 | /0737 |
Date | Maintenance Fee Events |
Nov 17 2009 | M1551: Payment of Maintenance Fee, 4th Year, Large Entity. |
Nov 23 2009 | ASPN: Payor Number Assigned. |
Nov 23 2009 | RMPN: Payer Number De-assigned. |
Nov 13 2013 | M1552: Payment of Maintenance Fee, 8th Year, Large Entity. |
Jan 15 2018 | REM: Maintenance Fee Reminder Mailed. |
Jul 02 2018 | EXP: Patent Expired for Failure to Pay Maintenance Fees. |
Date | Maintenance Schedule |
Jun 06 2009 | 4 years fee payment window open |
Dec 06 2009 | 6 months grace period start (w surcharge) |
Jun 06 2010 | patent expiry (for year 4) |
Jun 06 2012 | 2 years to revive unintentionally abandoned end. (for year 4) |
Jun 06 2013 | 8 years fee payment window open |
Dec 06 2013 | 6 months grace period start (w surcharge) |
Jun 06 2014 | patent expiry (for year 8) |
Jun 06 2016 | 2 years to revive unintentionally abandoned end. (for year 8) |
Jun 06 2017 | 12 years fee payment window open |
Dec 06 2017 | 6 months grace period start (w surcharge) |
Jun 06 2018 | patent expiry (for year 12) |
Jun 06 2020 | 2 years to revive unintentionally abandoned end. (for year 12) |